ARTICLE IN PRESS Available online at www.sciencedirect.com Research in Veterinary Science xxx (2008) xxx–xxx www.elsevier.com/locate/rvsc Evaluation of different API systems for identification of porcine Pasteurella multocida isolates Y.A. Vera Lizarazo, E.F. Rodrı´guez Ferri, C.B. Gutie´rrez Martı´n * Department of Animal Health, Section of Microbiology and Immunology, Faculty of Veterinary Medicine, University of Leo´n, 24007 Leo´n, Spain Accepted 5 February 2008 Abstract An exhaustive biochemical characterisation of 60 porcine Pasteurella multocida clinical isolates recovered from lesions indicative of pneumonia, previously confirmed by PCR and all belonging to the capsular serogroup A, was performed by means of four commercial systems. The API 20NE correctly identified almost all isolates (95%), but only 60% could be ascribed to this species by the API 20E method. The high diversity exhibited by the API 50CHB/E system, with six different patterns, does not advise its use as additional system for a definitive identification at the species level, but this method could be a potential tool for characterising P. multocida isolates below this level. The more uniform reactions yielded by the API ZYM test make this system helpful in the confirmatory identification of this organism. The high variability (20 profiles) obtained when the four systems are taken together also suggests their usefulness for epide- miological purposes in order to sub-type P. multocida isolates. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Pasteurella multocida; Swine; Biochemical characterisation; Identification Although Pasteurella multocida is part of the commensal In this study, we report an exhaustive biochemical char- flora of the upper respiratory tract of swine, it may induce acterisation of P. multocida clinical isolates recovered from pneumonia in growing and finishing pigs to the extent that porcine lungs, and evaluate the accuracy of API 20E in pneumonic pasteurellosis is one of the most common dis- their identification, as well as the capacity of API eases of intensively housed pigs. The disease is often asso- 50CHB/E and API ZYM as additional systems for a con- ciated with other important bacterial respiratory infections firmatory identification. The API 20NE, used as routine (enzootic pneumonia caused by Mycoplasma hyopneumo- system in clinical laboratories, was also included. niae and pleuropneumonia caused by Actinobacillus pleuro- A total of 60 P. multocida isolates were recovered from pneumoniae), and results in a disease condition described as swine with lesions indicative of pneumonia coming from 23 chronic respiratory disease. In this disorder, P. multocida farms in northeastern, northwestern and central Spain usually acts as a secondary pathogen, invading lungs between 2003 and 2004. The bacteria were first isolated injured by other organisms (Pijoan, 1992). In contrast, sys- on Columbia agar with 5% sheep blood (bioMe´rieux, temic pasteurellosis is not a frequently diagnosed disease in Lyon, France), and the mucous non-haemolitic colonies pigs, although outbreaks of acute septicaemia have been were further plated on a selective medium for P. multocida reported (Blackall et al., 2000). In addition, P. multocida (Avril et al., 1990): Mu¨eller-Hinton agar (Pronadisa, may induce progressive atrophic rhinitis in association with Spain) with 5% defibrinated horse blood (bioMe´rieux), Bordetella bronchiseptica (De Jong, 1992). amikacin (2 mg/l) (Sigma Chemical Co., St. Louis, Mo), vancomycin (4 mg/l) (Sigma), and amphotericin B (4 mg/ l) (Sigma). The isolates were genetically confirmed by use * Corresponding author. Tel.: +34 987 291 203; fax: +34 987 291 304. of a multiplex PCR assay (Townsend et al., 2001), and E-mail address: [email protected] (C.B. Gutie´rrez Martı´n). all of them were found to belong to capsular serogroup A. 0034-5288/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.rvsc.2008.02.002 Please cite this article in press as: Vera Lizarazo, Y.A. et al., Evaluation of different API systems for identification of porcine ..., Res. Vet. Sci. (2008), doi:10.1016/j.rvsc.2008.02.002 ARTICLE IN PRESS 2 Y.A. Vera Lizarazo et al. / Research in Veterinary Science xxx (2008) xxx–xxx Each clinical isolate was processed using the API 20E, excellent identification, with a 99.9% likelihood level and API 20NE, API 50CHB/E and API ZYM systems (bio- without tests against). Pattern II (0100004) was identified Me´rieux), along with the P. multocida NCTC 10322T refer- as either P. multocida or Pasteurella spp. (identities of ence strain. In addition, control strains were used as 3.8% and 58.4%, respectively); in addition, three other recommended by the manufacturer (Klebsiella pneumoniae organisms were proposed (Pseudomonas cepacia, Shewanel- subsp. pneumoniae ATCC 35657 for API 20E and API la putrefaciens and Moraxella spp., with identity rates of 50CHB/E, Aeromonas hydrophila ATCC 35654 for API 12.0%, 10.2% and 9.6%, respectively), and the API 20E 20NE, and Bacteroides thetaiotaomicron ATCC 8492 for database indicated that the profile identification was not API ZYM). Each P. multocida isolate was inoculated into valid for the 16 isolates comprised in this pattern. Finally, 0.85% NaCl, and turbidity was adjusted to 0.5 MacFarland the remaining eight isolates (pattern III: 0040004) were standard for API 20E and API 20NE; to 2 for API 50CHB/ misidentified as Vibrio hollisae, the profile identification E, and to 5 for API ZYM. The inoculum was distributed being considered as acceptable. In our study, a phenotypic into test strips according to the manufacturer’s instruc- diversity concerning ornithine decarboxylase, indole, and tions. API ZYM strips were read after a 4-h incubation fermentation of mannitol and sucrose has been obtained at 37 °C; API 20E strips after a 24-h incubation at 37 °C; (Table 1), all being characters that have been considered and API 20NE and API 50CHB/E strips after a 48-h incu- as essential for the identification of P. multocida at the spe- bation at 30 °C and 37 °C, respectively. Biochemical reac- cies level (Larivie`re et al., 1992). However, two recent stud- tions were read as positive or negative, translated into ies have demonstrated that several isolates from diverse numerical profiles (for API 20E and API20NE), and inter- origins that had been regarded atypical for these four phe- preted with the software APILAB Plus update 3.2.2. For notypic characters, they were all ascribed genetically to P. API ZYM, the reactions were considered only positive multocida by 16S rRNA gene sequencing and DNA–DNA for a strong activity (P30 nm of hydrolysed substrate), hybridization (Christensen et al., 2004, 2005). according to the recommendations of the manufacturer. The biochemical reactivity of the P. multocida isolates The four commercial systems were tested for their repro- examined here was quite different from the API 20E results ducibility, by use of P. multocida NCTC 10322T, which previously reported for 14 paired isolates recovered from was replicated five times, as previously described (Purdy the lungs and kidneys of the same slaughtered pigs in Den- et al., 1997). mark and Canada, all giving an identical pattern (0144524; The API 20E profiles allowed us to establish three phe- excellent identification, identity of 99.9%) (Buttenschøn notypic patterns for the P. multocida isolates (Table 1). The and Rosendal, 1990). However, our study was focussed profile with the highest number of them was pattern I on the evaluation of the identification capacity of the (60%), followed by patterns II (26.7%) and III (13.3%). API 20E system at the species level, while that of Butt- Pattern I also included the reference strain NCTC enschøn and Rosendal (1990) was undertaken to evaluate 10322T. The API 20E strips correctly identified at the spe- this system with sub-typing purposes. On the other hand, cies level only the isolates belonging to pattern I (0040524; our results closely agreed with those obtained by Collins et al. (1981), with only 64% of 50 strains of P. multocida isolated from different animals being accurately identified Table 1 Biochemical patterns of 60 porcine clinical isolates of Pasteurella by the API 20E system. In view of our results, the use of multocida yielded by the API 20E system API 20E for identification of this organism at the species Pattern level is considered unsatisfactory, and the widespread use of this commercial system in the routine identification in I II III clinical laboratories might therefore lead to misidentifi- Ornithine decarboxylase À + À cation of porcine P. multocida infection. Indole production + À + Fermentation of + ÀÀ Concerning the API 20NE strips, 57 field isolates (95%) mannitol, sorbitol and and the reference strain did not assimilate glucose and man- sucrose nitol, and their numerical code (3000004) allowed us to Reduction of nitrates + À + ascribe them to P. multocida, the identification level being No. (%) of isolates 36 (60.0) 16 (26.7) 8 (13.3) good (accuracy of 93.1%). In contrast, the remaining three API 20E profile 0040524 0100004 0040004 Result and likelihood Pasteurella Pasteurella spp. Vibrio isolates were positive to the two aforementioned tests level (%) multocida (58.4) hollisae (3044004), and the second pattern obtained in the API (99.9) Pseudomonas (80.6) 20NE system yielded a low discrimination in the correct cepacia (12.0) identification of this species, with an accuracy of only Shewanella 42.9% for P. multocida. Besides, these isolates were also pro- putrefaciens (10.2) posed to be representatives of Aeromonas salmonicida (iden- Moraxella spp. tity of 55.2%). However, these three isolates were correctly (9.6) ascribed to P. multocida by the API 20E system (pattern I). Pasteurella The API 20NE system has been developed for gram- multocida (3.8) negative organisms other than Enterobacteriaceae from Please cite this article in press as: Vera Lizarazo, Y.A.